specutils.manipulation gaussian_smooth does not work on numpy.int64 input stddev

[Rplesha]

I was trying to loop over a range of smoothing factors, which resulted in a confusing error.

Example code:

for factor in np.arange(11, 21, 2):
    smooth_data = gaussian_smooth(spec1d, stddev=factor)

resulting in the error:

ValueError                                Traceback (most recent call last)
Cell In[25], line 2
      1 for factor in np.arange(11, 21, 2):
----> 2     smooth_data = gaussian_smooth(spec1d, stddev=factor)

File ~/miniconda3/envs/jwst/lib/python3.11/site-packages/specutils/manipulation/smoothing.py:185, in gaussian_smooth(spectrum, stddev)
    183 # Parameter checks
    184 if not isinstance(stddev, (int, float)) or stddev <= 0:
--> 185     raise ValueError("The stddev parameter, {}, must be a number greater "
    186                      "than 0".format(stddev))
    188 # Create the gaussian kernel
    189 gaussian_kernel = convolution.Gaussian1DKernel(stddev)

ValueError: The stddev parameter, 11, must be a number greater than 0

There are of course multiple ways to get around this, including changing to np.arange(11.0, 21, 2), or creating a list of numbers, but I don't see a reason why int64 should not be included as an acceptable type. If there is a valid reason though, I suggest adding a more pointed error message.

I believe the relevant code to update is at:

specutils/specutils/manipulation/smoothing.py

Line 184 in 6dcc1ad

if not isinstance(stddev, (int, float)) or stddev <= 0:

I'm running specutils v1.19.0 on a mac running Sequoia 15.3.2, python 3.11.7. spec1d in this instance is a JWST NIRSpec dataset.